The phased array antenna is one of the most advanced radar technologies which allows multiple bean pointing and fast non-mechanical steering of microwave beams. The technology has promise for broad-band (2-20 GHz) free-space radar communications that can be used for a variety of commercial and military applications. Beam pointing/steering control systems are known, including true time delay systems and phase shift systems, for phased array antenna, true time delay systems being preferable since the steered beam angle is independent of frequency and squint is eliminated.
For a given microwave frequency f and number of microwave radiating elements along one direction N, the maximum time delay .DELTA.t.sub.max required to steer a beam over .+-.90.degree. is given by N/f. Also, the minimum temporal resolution .DELTA..tau..sub.min to achieve resolution R is given by 1/(f.multidot.R). Assuming a frequency range of 2-20 GHz, N=100 and R=1,000, .DELTA.t.sub.max =5-50 nsec and .DELTA..tau..sub.min =0.05-0.5 psec.
In conventional electronic RF systems, true time delay is achieved using switched lengths of electrical waveguide or cable. Such devices tend to be bulky, expensive, have high loss at high frequencies, and are susceptible to electrical crosstalk (due to electromagnetic interference) and temperature induced time delay changes. Recent advances in photonic technology can provide a better implementation of true time delay due to a natural high parallelism and large bandwidth as well as immunity to electromagnetic interference.
Heretofore known or suggested photonic true time delay systems have been configured so that each microwave element requires R fixed time delay generators, R switches and an R to 1 combiner. Thus, for a two dimensional (2-D) array with N.sup.2 elements in such systems, N.sup.2 R time delays and N.sup.2 R switches have been required. The insertion loss is mainly determined by the R to 1 combiner and is given by 10 log.sub.10 R. Although such a system is capable of adaptive beam forming as well as beam steering, it requires a tremendous amount of complexity, making its hardware implementation extremely difficult.
Although this complexity can be reduced to some degree by free-space path-switching methods, this still requires a cascaded array of many independent time-delay generators and parallel (N.sup.2) switches in 2-D spatial light modulators. Moreover, thus configured, the system presents other limitations, such as speed and path-dependent insertion loss.
A highly dispersive fiber prism method has been suggested and/or utilized that can significantly reduce the complexity as described above. However, this method requires very long (20 km for 1 GHz), N.sup.2 fiber bundles and a fast tunable narrow linewidth light source with broad tuning range. It has been suggested that the long length could be significantly reduced by using an array of fiber gratings, but significant problems with this implementation would yet be posed. Most of the heretofore suggested approaches for use of fiber gratings as a means to generate true time delays employ an array of normal single frequency fiber gratings, the desired time delays being selected by a tunable narrow linewidth light source. To achieve high resolution, both a broad tuning range and a narrow linewidth are required. Moreover, the wavelength would need to be changeable rapidly (within a few microseconds--a speed unattainable by current laser technology) for effective implementation.
In addition, two dimensional (2-D) extension architecture for such photonic true time delay systems as have been heretofore suggested could utilize further improvements. Conventional image rotation has been accomplished, for example, by rotating a dove prism by an angle .theta. around the optical axis, the output image thus being rotated by 2.theta.. Such conventional rotation thus requires mechanical movement of components and is, therefore, inherently slow and lacking adequate unreliability.